Mendelian randomization combined with single-cell sequencing analysis revealed prognostic genes related to myeloid cell differentiation in prostate cancer and experimental verification.

BACKGROUND

Myeloid cell differentiation (MCD) has an important correlation with prostate cancer (PCa), but the mechanism of action of the former in the latter is still under investigation. This study designed to investigate the prognostic genes related to MCD in PCa and the associated mechanisms.

METHODS

The related data were downloaded from public databases. Differentially expressed genes (DEGs) were intersected with MCD related genes (MCDRGs) to acquire candidate genes. Candidate prognostic genes with a causal relationship to PCa were further obtained through Mendelian randomization (MR). Prognostic genes were acquired by univariate Cox regression analysis and Least Absolute Shrinkage and Selection Operator (LASSO) analysis. Then, the risk model was built based on prognostic genes. Immune infiltration, nomogram model, and drug sensitivity were employed to investigate the roles of prognostic genes in PCa. The manifestation of prognostic genes in key cells was also investigated by single-cell sequencing (scRNA-seq) analysis. Finally, the manifestation of prognostic genes were authenticated by in vitro experiments.

RESULTS

The 23 candidate prognostic genes had a causal relationship with PCa. The 5 prognostic genes (NR3C1, BMP2, RACGAP1, TLR3, FASN) were identified. The risk models suggested that high risk group (HRG)'s survival rate was inferior to that of low risk group (LRG). The nomogram indicated that prognostic genes could effectively predict the survival status of PCa patients. There were 18 immune cells that suggested notable differences between the HRG and the LRG. The HRG and LRG suggested notable differences in sensitivity to 86 drugs such as AZD8186. Epithelial cells were considered as key cells. Only FASN was consistently active during critical cell differentiation. The in vitro results were consistent with the results of bioinformatics analysis, indicating that the analysis results were reliable.

CONCLUSION

This study identified 5 prognostic genes and a risk model, suggesting a fresh thought on the subsequent development of PCa related drugs.